ret2reg

Using Registers to bypass ASLR

ret2reg simply involves jumping to register addresses rather than hardcoded addresses, much like Using RSP for Shellcode. For example, you may find RAX always points at your buffer when the ret is executed, so you could utilise a call rax or jmp rax to continue from there.

The reason RAX is the most common for this technique is that, by convention, the return value of a function is stored in RAX. For example, take the following basic code:

#include <stdio.h>

int test() {
    return 0xdeadbeef;
}

int main() {
    test();
    return 0;
}

If we compile and disassemble the function, we get this:

0x55ea94f68125      55             push rbp
0x55ea94f68126      4889e5         mov rbp, rsp
0x55ea94f68129      b8efbeadde     mov eax, 0xdeadbeef
0x55ea94f6812e      5d             pop rbp
0x55ea94f6812f      c3             ret

As you can see, the value 0xdeadbeef is being moved into EAX.